2013 NCWIT Summit - Plenary I, Building Meritocratic Communities by Michael Schwern and Nóirín Plunkett

August 14, 2013

[silence 0:00 to 0:05]

[music 0:06 to 0:14]

Michael Schwern:  Howdy.

Audience:  Howdy.

Michael:  Put that over to the side. I'm Michael Schwern, and I'm a special bunny. I get my own laptop, but I don't get it to be displayed, oh well.


Michael:  I'm a PERL programmer. I've been doing it for about 15 years, and people who don't know, PERL is an open source programming language. It's been around for about...Well, it was very popular in the dot‑com boom. If you don't know PERL, it's the one that looks like line noise, with all the dollar signs in it that you probably yell at. Yeah, there you go.

Let's review. [clears throat] I do a lot of work on the infrastructural language. When you install something that has to do with PERL, you probably go through about three subsystems that I've written, maintained or overhauled. The point of that is just like, I have some sort of influence in this community, I have some sort of power in that open source community.

Why am I here? I have no degree. I have no company. I have no university I'm associated with. I kind of looked through the speaker list and I went, "Whoa, I'm way out of my league here. I'm not CEO of anything."


Michael:  I'm getting a sort of first hand of imposter syndrome, certainly.


Michael:  I've had no formal training in sociology or gender politics. I'm a total noob, I don't even have a software degree. I failed out of one of the finest institutions in the country.


Michael:  What I want to say is, I'm relatively new to all this and take what I say with a grain of salt. These are more about my personal observations. I'm here because I gave a keynote last year at YAPC North America, which is basically the PERL conference in North America about diversity.

I was asked to give a keynote and that's what I decided to do it on, using the analogy of "Star Trek" captain's command styles. Jenny contacted me and she said it charmed their socks off. I'd never heard of NCWIT before, so here I am. This is going to be a talk about that talk...


Michael:  ...and how it was written. It gets worse. How it was written...


Michael:  ...and the particular audience it was written for, also the consequences of speaking out in that audience. I'm also here because I'm a guy, which is kind of a weird thing to say at a women in IT audience, but a guy giving a diversity keynote at an open source conference is really rare. What is interesting about doing it is that I still can have the old perspectives of speaking their language, understanding their point of view. It's starting to fade, but it's still there.

I can have some emotional distance from the problem, right? I haven't been discriminated against in that way. I can speak a bit calmer. I can take a bit more punishment, particularly in public. I can use my power and privilege for good. I'm not going to get seriously harassed. I'm not in danger. That's fading fast. That's part of what I'm going to be talking about.

Guys will listen to guys. I'm very safe. I don't look like some wild feminist. I'm a beardy bloke. I found this out at another Open Source Conference about two or three years ago, when I publicly withdrew from speaking to emphasize their lack of Code of Conduct, and their lack of dealing with incidents. But I still attended, and all the guys were coming up to me and were like, "What's going on? Why? Why? Is there a problem? What's the Code of Conduct? Why do we need it? Tell me."

I was the safe person to talk to and it was really interesting. I learned a lot about how to talk to guys, particularly in Open Source, about all the issues with women in IT, and particularly what their viewpoint was, and what their gaps in learning, and what their concerns were. Now unfortunately as an indirect result of giving that talk, I'm not going to the PERL Conference this year. I have withdrawn my talks and I'm not attending, more on that later. This just happened.

I kind of had to rapidly revise the talk late into the night. I'm a bit close to the problem right now, so please keep that in mind as I give this talk. You're going to get a sort of pretty bad view of Open Source.


Michael:  I apologize for that but it's a true view, but it's one of the worst they can throw at a guy anyway. Now, so I'm giving a keynote at a conference about how I wrote a keynote for another conference last year, and I'm not going to that conference this year, because of the keynote last year. It's going to be a little bit like "Inception."


Michael:  OK enough about me. Let's talk about you, by talking about me.


Michael:  The keynote that I gave was called PERL The Next Generation. That was the theme of the conference, The Next Generation. I figured, "Well, it's very interesting to be talking about the next generation of PERL. We should talk about diversity because that's a very generational thing." You're not an Open Source audience, as anybody can understand. Let me frame the problem for you.

Open source, in general, is overwhelmingly, guys. 95 percent is the usual number that comes out. For example, PERL has had 36 release managers over 20 years. Most of those have been in the last few years because they've been rotating faster. They learned you burn people out pretty fast, zero women out of 36, none what‑so‑ever.

Obviously, the 95 percent male, and they're all in power, is going to have a huge influence on anything, but this audience gets really uncomfortable when you talk about gender or demographics or diversity, really uncomfortable. Generally because, like me, they're complete and utter noobs about the subject, but they're the sort of people who think that because they are smart in one field, they can figure everything else out. They don't really need to do any research or learn anything. They tend to mix up...They're engineers, they want to solve a problem, but they can't solve the problem immediately.

They're not comfortable with saying, "I don't know," so they have a tendency to dismiss the problem, and say, "You don't actually have a problem, I'm not acknowledging your problem," or "It's not a problem for me, therefore, it's not a problem for you." That's very difficult to deal with. There seem to have very libertarian views, and putting that all together makes discussing social issues really, really frustrating with this group, because they're all the same. It's very, very difficult to break in if you don't match that mold.

Why is it this way? Well, one of my hypotheses is because open source is almost entirely about software, computing, and IT, there's no counterbalance groups. There's no HR, there's no marketing, there's no executive creative department. There's no more traditionally diverse group to inject some diversity into the system. It's all, traditionally, been male dominated groups, and it continues to be that. Who knows why it is? Sorry, I'm sure people here know why it is.


Michael:  But, it definitely is, there's no denying that. Mind you, I'm describing the group in power. I'm describing the groups that have voice and they have control of the center. I'm generalizing quite a bit about this group. In particular, I'm drawing my experiences from PERL, which is kind of old‑school, and we're a language, so we kind of tend to be even more old‑school than the other communities.

Those that are not in power actually look rather different. They're far more heterogeneous. They're the users and the minor contributors and the projects that are more directly about the users and not about the technology. The further away from the centers power you get, the more diversity that you find.

There are also large groups in the middle like Wiki‑media and Mozilla, AID initiative, feminism, dream‑width, who are really addressing the problem and they're doing a great job. All of this I'm talking about is based on my experience in PERL. There are fantastic people out there, and what's with all the ponies, seriously? The community looks something like this. There are some real jerks.


Michael:  On the other side there are some people who get it, and the middle, they mostly just want to code. They don't really care about social issues. They just want to be left alone. Occasionally they're jerks, but on the whole they're good people. I really do like most people in PERL and open source. Overwhelmingly, I like them.

Unfortunately, because open source is generally un‑moderated it feels like this. This is what you hear and what you see, because you gain power and influence in open source by speaking and writing on mailing lists and in chat channels. The occasional outburst is un‑moderated, they're not taken care of, so they have undue influence on everybody else. The jerks get an undue amount of power, and the good people tend to be shouted down, and they just leave, because why would you want to deal with that?

It's sort of rule by the shoutiest, and jerks like to shout. The big question that I had to ask about this talk is how do you give a talk about diversity to an audience who isn't diverse? How do you give a talk about diversity to an audience who basically feels, thinks, and acts the same in certain ways?

They think they're very happy with the way things are. They think that everyone else is very happy with the way things are. They hold all the power, they hold all the voice, and they see diversity issues as completely optional, and they get very upset when you talk about diversity to them. They get very upset when you point it out.

For example of what the kind of people you have to deal with is, the group will almost entirely look at diversity in terms of how it will impact them. In terms of, if you have a code of conduct or any kind of anti‑harassment protection. For example, this is a response that came up in a conversation about hugging. They actually have a hugging protocol, it's that dysfunctional.


Michael:  Somebody basically expressed that people shouldn't touch me without asking and that includes hugging. You can just ask. It's OK. The response from one of the guys was, if anyone so much attempts to come...I'm sorry, this was his caricature of somebody's response to being hugged without asking, is that they were going to scream assault, invoke the Code of Conduct, and do whatever necessary to stomp out the perpetrator. This is purely a punitive thing against him, just for trying to be nice.

That's the viewpoint, the reversal of these protections is, "How is it going to affect the people in power?" The horrible thing is that this is not considered bizarre. This is totally within the realm of normal conversation. It's kind of on the far end of the edge of the window, but the window of accepted discussion is so far...it's like in another building, you can't even find it.

It's amazing when you read these discussions and that's kind of scary. I was talking with somebody on twitter this evening who told me that the reason that Open Source is 95 percent men is because the women just aren't interested. The scary thing is, I read that and I went, "You know, people are going to listen to you. That's really terrifying."

As I was handing in reports and papers and studies and being like, "No it's really not that." [laughs] Yeah, conversations quickly go to that point. The first step in this keynote was convincing the guys that the problem even exists. The simple problem, gender imbalance, you have a gender imbalance in your community. They don't think about these things despite it being all around them, right in front of their face at all times, especially at the conference.

They probably don't realize how bad the gender imbalance is, because they just don't have to think about it. It's sort of a definition of privilege, "I don't have to think about it." One of the first things I like to do at conferences is, I get all the guys, all the people I identify as guys, to stand up. It would be a very different experience in this audience. Either stand up, or raise their hand if they can't stand up and just look around. It's right there in front of your face. You can't deny it, 95 percent of men at the conference.

It's undeniable and it's immediate. It's a way to kind of just put it right in front of their face so they can't ignore it. Cut through the privilege of having to talk about privilege. Next, once I've been like, "There's a problem," convince them that it is their problem.

A very common excuse is that, "It's not Open Source, it's something else. It's not computing it's something else." "The women just aren't interested," is kind of an extreme example of that. People like to talk about, "Oh there's just no women in computing. We're just a reflection of that. It's not our fault."

It's an excuse for inaction. It's not their problem. They can't do anything about it. National Science Foundation says that women make up about 28 percent of proprietary software development. In the Open Source it's single digits. Most Open Source developers don't know this. It's always a surprise when they find this out.

These are the numbers for PERL Conference Europe, was six percent of women attending. I don't know what their more recent numbers are. They stopped reporting gender.


Michael:  They report age and other demographic information. No gender. The speaker list for the PERL Conference last year looked like this. The highlighted people are the women speakers. That's another thing they like to show them. It's like right in front of your, There's a problem.

Again single percentages, I think that's 7 out of 100. This is again a common pattern across most of Open Source. I like to use these charts. I like to use this data because first of all it comes from an authority. It comes from the National Science Foundation. It comes from a paper commissioned by the European Union as well, and they can't really be argued with.

There is something about Open Source that is not including women. The data that comes from this site called flosspols, [www.flosspols.org] if you're interested more on the topic about inclusion of women in Open Source, they have two fantastic papers. They're right on the front page. They lay everything out about what's wrong and how to fix it.

It was commissioned by the European Union, by the University of Maastricht, sorry anybody whose city I've been butchered, University of Cambridge, back in 2006. Open Source people have not read this document. It's pretty embarrassing. They lay out everything, "D16" is the major paper. They lay out everything in the first section, the key findings section. "Open Source participants view technology as an autonomous field separate from people..."


Michael:  "...this means that anything they interpret as social is easily dismissed as artificial social conditioning. Because the conditioning is considered more less arbitrary, in their view it is supposed to be easily cast aside by individuals choosing to ignore it." So if you don't want to be discriminated against you can just not be discriminated against.


Michael:  "Open Source also has a deep..." Sorry FLOSS is Free/Libre Open‑Source, "...has deeply volunteerist ethos which values notions of individual autonomy and volition. As a result participants largely do not believe that gender has anything to do with their own individual action. The situation is thereby perpetuated in spite of the expressed desire for change."

That is essentially homogeny breeding homogeny, and in‑action combined with homogeny, perpetuating homogeny. Homogeny? Homogeneity? Oh well, sorry. Did they mention I'm new to this? [clears throat] Then it goes on to list ways that women are actively excluded by Open Source projects.

We value coding over results, so any work on documentation or design or management or anything else, doesn't get the chops as just pounding away at code. There's an assumption that you have a long history with computers, you already know how to use them, you know what a text editor is, a command like is, you know how to use Unix, all this kind of stuff. You understand what a program is, and algorithms.

We tend to skip over that part when we're on‑boarding new people, if we bother to on‑board new people. There's an acceptance of verbal abuse in "Flame Wars," that a lot of people just don't go for. I'm really getting sick of it, and that there's a need for long hours of leisure time to devote to Open Source. People think that in Open Source, if you can't devote 20 hours a week, you're worthless.

You can't just put in a few hours week, and all the systems are built around that assumption. Based on this bit of data I go on to state this to the audience, a very carefully worded statement, "There is something about Open Source which does not include women." This is very particularly crafted for that audience of 95 percent men who do not want to hear about diversity. Note, I did not use a word like "problem," or "wrong," "This is wrong, this is a problem," and I carefully did not say, "Open Source drives away women."

The goal here is to merely get folks to agree that the problem exists. Acknowledge the current state without having to admit that there's a problem, just get them to admit the current reality. It's important fiction to move things forward. It's very baby steps. I had trouble with even this mild a statement. That's kind of those baby steps going forward, Even if we're not...you know, Etsy did some fantastic work on increasing their engineering diversity by creating a hacker school and hiring more junior programmers.

You can find out more about that at https://bitly.com/etsygrowth, I'm going to skip forward. OK, so step two, now that we've accepted the problem, we need to explain how it happens, but in a way that doesn't trigger this very, very touchy audience. If they feel guilty or blamed, or like their bad people, they can react very poorly. The first rule of privilege is you don't talk about privilege.


Michael:  The second rule of privilege is you don't talk about privilege. That's part of the definition of privilege. You don't need to talk about it. My rule of thumb is if you bring up privilege to male privileged, or an open source audience, either your talk is done, because nobody's listening to you anymore, or you are talking about nothing but privilege, explaining the concept of privilege, for the entire rest of your talk.

While I'd love to do that, that wasn't my goal last year, so I came up with something else that kind of has similar effect, and that was homogeneity. The idea that open source is the way it is because it's demographically and socially homogenous. Is this true? It's a nice story. It's probably true, but it's really useful.

It gives people an explanation about why a group of good people, that can remain good people, I am a good person, can do bad things. Individually, these individuals are working to optimize their group to make it happier for people around them, their friends, but this creates an unwelcoming environment for people who are outside of the group. It's very convenient excuse. It gets them to actually, again, continue to listen to you because you got to deal with this mass of people who don't understand what privilege is, and don't care, and get them to not fight you.

Given my experience this year, this may not have been such a good idea. We'll find out. The second thing of that is, now, I tried to move the focus away from women. Because gender is a really touchy topic and you don't want the few women around to be barraged, and again, you don't want to trigger all these guys from the guilt about the gender politics, gender diversity, feminism, so on and so forth.

I said, "Gender is very easy to track," so, "demographic diversity is the canary in the coal mine for greater diversity of the group." If gender's out of whack, what else is out of whack? It turns out it's totally true. One of the examples we had is, "Let's talk about operating systems." That's never a problem at all on an open source environment.


Michael:  The people who do the corporate PERL training got together, and they said, "You know, our students are about a 50‑50 split between Windows and UNIX." These are people who are outside the PERL community, the inner center community. The core community is something more like 90‑20.

Hmm? Right there, there's something we're doing that's excluding Windows users, and you can do this other attributes, again, people who don't want to have flame wars, people who...the number of hours you can devote a week, so it's not just about women, it's OK, you're off the hook, whew, to the guys.

But these are totally things that also turn out to exclude women, so it kind of serves its purpose, but it's a story. It's a nice bedtime story to get them do what I need them to do. This gets into, what I call, "value alignment," there's probably an existing term for it. Part of the realization was that people either value diversity or they don't.

In a key note, I'm not going to change their minds on that one. I needed to use some other way of doing it, and basically, I need to explain to them why diversity is valuable to them according to their values, and show that they're not in conflict, and it's not to zero sum game. "It's OK guys, you don't have to be terrified that you're going to lose all your power and privilege just because different people are going to enter the group."

Their values are things like community growth, especially in a language like PERL that is considered to be waning. They're concerned about the size of the community. Well, this becomes a no‑brainer. 25 percent of your potential people are being dropped in the floor. You're losing that many potential from the pool of programmers. Hey, just get a huge boost to the community, "Be welcoming to Windows and double your size." It's a very simple sell.

Another one is the concept of the meritocracy, the concept that you gain power through your own personal merit. This is a core value of open source. The argument that I use is very simple, "If open source is five percent women, and commercial software is 30 percent, then meritocracy isn't working, it's broken." This is ideas of fairness, and ideas about libertarian values, and fixing diversity fixes the meritocracy.

The other one is basically getting people to fill and don technical roles. A viewpoint of the way that open source is laid out is it's basically completely designed for system administrators and developers, and not so much for writers, and designers, and project managers, and users, and everybody else. We have serious trouble finding anybody who is not a system administrator or developer to join in.

Well, if you make yourself more welcoming to people, you will start to get these roles filled that we traditionally don't get filled. What sorts of people are we driving away? What ideas are we losing? What sort of viewpoints and skills are we losing? At that point, I've laid out the problem...


Michael:  ...and that is their problem, and that is beneficial to solve that problem. But it just won't go away. It will just get worse if you do nothing about it because homogeneity combined with a group that doesn't really study social issues breeds more homogeneity. When I was developing the key note, I talked to people, and a common theme was, "We don't want to just hear about the problem anymore. Give us solutions." OK, so we have a group of thick skinned, socially passive, libertarian men. How are they going to solve this problem?


Michael:  You got to make it fun and interesting because, again, they really don't want to hear about this topic. Normally, one of the reasons I can give the talk is because I'm a guy. They'll listen to a guy about this. They won't just tune it out. It was a key note. They had no choice, just like you.


Michael:  You have to speak their language. That's where "Star Trek" comes in. Apologies to anybody who's not a Star Trek fan in this audience. I'll try and make sure everything is explained. Specifically, I was talking about Starfleet captains and their command styles. I want to take a quick poll. Raise your hand if you're a Kirk fan, old Kirk, not fake‑y new Kirk. Picard? Sisko? A little less, Janeway? Porthos?


Michael:  OK, yeah. The real brains behind Enterprise for sure. I want talk about the two most iconic captains which are Kirk and Picard. They are two very different Starship captains, 25 years apart, each from a different generation, and don't worry if you don't know Star Trek. Again, I'll explain.

Captain Kirk, from the original '60s show, captain of the classic USS Enterprise, NCC‑1701, no A, B, C or D. When Kirk is in a crisis, Kirk takes action. Kirk gives the orders. Kirk is decisive. Kirk beams down to the planet. Kirk punches the aliens. Kirk kisses the girl. It's Kirk, Kirk, Kirk, Kirk, Kirk, Kirk, Kirk. Sometimes he gets advice from his two best buds, but otherwise, it's all about Captain Kirk.


Michael:  It's Captain Kirk's ships, and that's great.

[laughter] [applause]

Michael:  That's great for Kirk and his crew. Captain Picard from the early '90s, "Star Trek: Next Generation," captain of a much newer, bigger Starship Enterprise, much more glowy. When Kirk [sic] [Picard] faces a crisis, he calls a meeting. [clears throat] He gathers his senior staff together, and he discusses the problem, and he gets their opinion, and they express them. Worf, the security officer, wants to shoot it. Geordi, the engineer, wants to study it. Beverly, the doctor, wants to know if it's hurt. Riker, the first officer, wants to have sex with it.


Michael:  Troi, the ship's councilor, wants to know how you feel about it. Data, an android, feels bad because he doesn't have feelings.


Michael:  Picard listens to all their different viewpoints, and they're there at the source of the crisis, and they have power, and they have voice, and they're empowered to help him make an informed decision, and then they have staring contest.


Michael:  The difference between Kirk and Picard's command style is what I want to point out. Kirk is all about Kirk. Kirk is all about doing it his way. Picard is all about collaboration. In PERL, and open source in general, is almost entirely made up of Kirks, we're almost...The term is benevolent dictator.


Michael:  Benevolent‑ish dictator. Whoever started the project controls the project, and dictates how the community will act, and forever and ever, amen. I control my projects because I started them, not because I'm the best person to administrate them. Picard builds a senior staff. It's made up of multiple people and multiple viewpoints.

He's empowering people, not having power over people. We had Kirk's for so long. If we get some Picard's, maybe we have some Sisko's and some Janeway's. I'll admit it, I'm a Kirk, totally a Kirk. I'd love to be a Picard. I don't want to just shave my head and call it good.


Michael:  A few things that I discovered is, again, I wanted to figure out solutions that people can do in their projects that are concrete. It's not just like, "You should be more diverse." It's like, "Well, what can you do about that?"

One of the concrete ways diversity can be increased in open source is anything you do to make it easier for new contributors and new users, anything you help for new user outreach, is going to improve your diversity because they can't possibly be any less diverse than the way at which you are right now, and bringing in the next generation, and letting them use it.

In this example, I finally wrote down the release process for my projects, so other people could do it. It was just all in my head. It was kind of in the collective consciousness, and then got people who were new to the project, and they were like...You know, if somebody emailed me, and were like, "Hey, there's this bug that I fixed. I really needed a release of it."

I'd be like, "You want to do it?" They're like, "OK," and we step through it, and fix up the process, and iterate to the point where I can pretty much run somebody who's new to PERL through the process at this point, which is great. It's dealing with, what I like to call, "a step zero problem." The problems that we don't consider, the experts, to be information that people need to know about. Something we totally take for granted. People afterwards pointed out that it's a bit awkward to be key note about diversity and hold up to old white guys as models.


Michael:  That was a point that I didn't make strongly enough in the key note at the time, but the point is that power in open source is vested in white guys, and they have the power to make the change happen, and happen fast. If it's awkward to talk about it this way, it's only because the reality of open source is awkward. It's hard to face that awkwardness, so you make it a little bit funny and make them go, "Wait a second. Oh, I get it." Yeah, it's going to be awkward.


Michael:  They're going to get it horribly, horribly wrong. They are getting it horribly, horribly wrong, but it's advancing, and the first season is always awkward, and you learn by doing and by failing and trying again. More on that a little bit. I want to give a shout out to Fashion It So, which is the blog that I took all this from, it's a fashion blog about Star Trek: Next Generation, they have all these screen caps. It's hilarious.

I also want to give a shout out to...One of the reason I became convinced about this "fix diversity at the top, and it'll all be much more...easier to deal with throughout the entire process," was by a conference in my hometown called Open Source Bridge, an open source conference that's about...Well, it's open source, but it's explicitly about open source citizenship, about all the people involved in open source, not just developers, but the users, the admins, the business people, institutions, designers, journalists, newbies, and it's all about what you're doing.

It's more about what you're doing, value in what you're doing, not just the code. That right there fixes one of the key problems that was pointed out in FLOSS polls. They got 25 percent women speakers, which is fantastic in open source. It gets us up to that, at least the level of proprietary software, so it's sort of eliminated at least the open‑source layer of bias going on. I think Linux‑conf Australia also hit 25 percent women through their efforts.

It looks very different, or for you, it probably looks very normal to have an open source conference, and it's not just all beardy blokes. It's no big thing, because it was all kind of set up and taken care of from the beginning. People had their viewpoints expressed during the planning process, not at the end, and having to have a big fight about it. How much time do I have? OK.

That's mostly what I want to say about the keynote itself, now there's an epilogue. I said at the beginning, I'm not going to the PERL conference this year, I withdrew my talk, and I'm not attending. To make a very long story short, the organization was more of a mess than usual this year, because, hey, no conference organizers in open source, just developers trying to run conferences.

I introduced three volunteers to help out, a budding a project manager, and PERL programmer who always felt he was a bit outside the community, wanted help out, and experienced open source conference organizer, who was also a previous speaker, and an event planner and volunteer coordinator who was not a programmer at all. Like, "You guys have organizational problems, here's the team to fix it, and they really want to help out."

It turns out there was no process to onboard volunteers, so they were just sort of dumped on the mailing list, and there was no process to delegate authority, there was no process to say, "Oh, you want to be a volunteer coordinator? OK, you're the volunteer coordinator, and I'm going to empower you, and here's what you're supposed to do." It was all just sort of loosey‑goosey, and if you didn't already have power in the PERL community, you could not get anything done in the conference.

The two things were not separated from each other. There was a huge split in communications. New people were all directed out of mailing lists, all the old guard talked on the chat channel that they never told anybody about. Then when we pointed this out, they said that, "Yeah, we make all of the decisions on the chat channel, and then we inform the mailing list," but they never even inform the mailing list.

We only figured this out at the end, after all the disasters had happened. The organizer didn't believe that there was a need to have anti‑harassment procedures, or people trained to deal with incidents. He just thought, "We have a code of conduct, were done. It's too much work, it's optional." Again, that kind of like, "It's not a problem for me. Therefore it's not going to be a problem."

There were incidents. There were incidents during the planning process, the conference hasn't even happened yet. Minor, but pretty clear verbal abuse, particularly on the conference chat channel, which is run by a totally different organization. The server admin's were involved. The chat admin's were part of the people giving out the abuse. I tried to talk to them, and got more abuse, didn't go well.

He said, "I'm only going to listen to the organizer," because again, no delegation of power. It was all just in the organizer, the only person they would listen to. When I showed the organizer he said, "Oh well, the admin was busy, you interrupted him," and sort of excusing this terrible, terrible behavior.

It's a pretty common excuse. I was told that this is a chat system issue, and that I should not be expected that the code of conduct applies outside the few days of the conference, and he took a very, very legalistic, and minimalistic view of dealing with incidents, which is again very, very common.

People in open source get very, very legalistic. It's that reversal of viewpoints of, "This is all about how it's going to affect me, the privileged, how it might restrict me, the work I might have to do about it, not about how it's going to protect people." That's a clue what I had been getting from the organizer. I escalated it to be an official code of conduct complaint, because that system was mostly in place, and the response I got back from the code of conduct committee was, "The scope of the code of conduct does not extend conversations on the chat channel that take place a month before the conference begins," the advertised chat channel.

Again, this purely legalistic, very narrow viewpoint, very minimalist point, very much like, "This is a burden, we don't want to deal with it," from the code of conduct committee, arguably this is there one job and they won't do it...


Michael:  ...and I wasn't even asking them to police the chat channel at all times, or anything. I gave them all the logs and I just said, "Get involved," and they wouldn't do it. The same people involved in that incident would be in control of the chat channels during the conference. It's not like there would be a big shifting of administration, or anything like that, so really what it comes down to is that code of conduct is a joke in this town.


Michael:  They left volunteers unsupported, and in danger. Nobody actually ever was like, "No, this isn't really in our scope, but since you brought this up, and since there was a volunteer having a problem, we will get involved this one time." The issue was just left on the floor. Supporting your volunteers should be a no‑brainer if you want volunteers.

It's kind of like calling up 911 because your house is burning down, and they are like, "Oh sorry, we only go to that neighborhood on weekends, call back then." [makes hang up sound] Then you have your neighbors all standing around your burning house going, "Yeah, I think it's a grease fire. Boy, we should really put together a volunteer fire department."


Michael:  That was the last straw, I withdrew. This is much better [laughs] than last year, which you'll see in a moment, and things are very, very much improving. But I think my tolerance for...can't say what I want to say. My tolerance for jerks and crap is lowering faster than they're getting their act together.


Michael:  I didn't just withdraw, I talked about it publicly. All my official channels were exhausted, so I figured the only thing I've got left is court of public opinion, which so I published on my own blog, to get it outside of the echo chamber and on Twitter, which I control. I've got like 1400 followers, which is three times the size of the conference.


Michael:  Now, I could do this because I am privileged, both within the PERL community, and in society at large, I'm a white guy, I'm pretty sure I'm not going to get serious physical danger, or harassment. Although I checked, I made sure that if my blog hit Gawker, I would be in trouble. I did a lot of preparation beforehand, hardening my systems, making sure I couldn't be hacked, and so on, and so forth.

I put it out there in the attempt to improve the sampling bias in the open‑source community. Last year, I want to talk about really fast, because I'm starting to run out of time, I had a problem with the organizer. A different organizer, because there's a different group running the conference every year, which is a problem, who posted a code of conduct, that said this, "All conferences treat each other with respect." What does disrespect mean? We obviously don't know.

"Call the police if you have a problem," thanks. If somebody's harassing you, ask them to stop. Then finally, after you've gone through all of that, you can report to the YAPC staff. I'm only going to explain to this audience why this is problematic, I tried to explain to him why it's problematic, and I got back things from him like, "How many people do you know who have been raped?"

I got back that I was going to badmouth the conference, that if I didn't like it, don't come. It was very clear that I threatened his authority. This is the person who invited me as keynote speaker, and I had never met him before. There was no pre‑existing animosity between us. When I brought it out other people...


Michael:  ...and said, "Hey, he's not listening to me, maybe he'll listen to you. He's not listening to guys. Maybe he'll listen to women." This email I got out of the blue from the organizer, "How dare you speak about your problems to other people?" It's a very classic verbally abusive pattern.

Now, he eventually relented, apologized, and put up a very good code of conduct, but it would be totally toothless. The point is, this is not the first time this has happened to me. Tolerance was very, very low this year. Now when I posted this information, I waited for the comments to roll in, and left them largely un‑moderated, and did not reply to them, because I wanted to find out a few things about the difference in the commentators on Twitter, versus on blog, and inside the community versus outside the community. I got about 20 insiders and 20 outsiders, not a huge number, but I've seen sociology papers with smaller samples.


Michael:  This really sucked, because I had to read the comments, and I was reading them last night and, "ugh," but it's for a good cause. I went over, after careful analysis with a team of doctors, here's what I got. What I observed is that insiders are far more likely to reply directly on the forum in question, than outsiders. Outsiders are far more likely reply in Twitter, and private email, and in tangential things, and not on the forum in question, where the topic is brought up.

That right there causes a dichotomy in what you see from the insiders versus the outside, and from on the forum in question as from the outside, and I've seen this, and this is the first time I ever measured it, very interesting. The other one I answered is that outsiders were overwhelmingly positive in their response. Now granted, this is all just what I decided was positive and negative, and who's inside and outside, but insiders were very mixed, half‑and‑half.

Again, this is a very, very different view of the problem. If you're on the inside, it seems like there's a big fight. If you're on the outside, it's a no‑brainer. The people who wanted to talk about the details of the incident, now mind you I did not talk about the incident, they brought it up. I talked only about the process failure. I specifically didn't talk about the incident because I didn't want to get any more hassle, and because it would distract from the point, the process failed.

The insiders wanted to talk, about half‑and‑half, much about the incident as the process failure, where as the outsiders totally got it, and were totally focused on the process failure. My conclusion, it's kind of like, "Why do I hang out with these people?"


Michael:  Yeah, I should be hanging out with people on the outside, so I've got some soul‑searching to do, and the community has some soul‑searching to do. This is an interesting measure of the echo chamber, and it's also an interesting measure of how a homogenous community enforces his homogeneity. They enforce it by, they talk on the main forms of medication, the insiders, they're the ones who have...they have a very different viewpoint about it than people on the outside, and they don't get it.

They don't really get the problem, the real problem. It will be interesting to see how the community reacts to this sort of data. I'm already preparing, "How are they going to explain this away?" Because they're really smart. When you don't want to solve a problem, and you're really smart, you figure out ways to dismiss the problem. OK, I think I've got some time left. Now at this point I want to hand off the mike to another speaker. Some of you may know Noirin Plunkett, she's very pink.


Michael:  She's here as an attendee in her own right. She's the only woman that sat on the board of the Apache foundation. She's an advisor to the ADA initiative, who are awesome. They're doing work for women in open computing. She's the director of the open cloud initiative, a computational linguistic, a technical writer, a dancer, now OK...you didn't take the cue.

She's been my mentor, and my support in dealing with these issues, because noob/expert, and she is also my wife.


Michael:  I wanted to bring up Noirin Plunkett.


Noirin Plunkett:  Thank you, Schwern says that he was invited here to speak because his next‑generation talk charmed the conference committee, and I hope you'll agree with me that he's an excellent speaker, but there is a deeper reason too. As a man, he has a very different perspective, and a different set of resources to try and combat the problem.

He's asked me to come up here and talk a little bit about my perspective as a woman, the backlash that I have faced in bringing up these issues and the tools and resources that I've found to combat it. Schwern has faced misunderstanding, and pressure, and some very rude emails from people he had previously called friends. Some of it's been awful, but it's been frankly quite mild in comparison to the reaction that me and my female friend space.

When I've spoken out without naming names, I've been accused of lying. I've been told I'm part of the problem, and that speaking out is making our industry look bad. I've had my concerns minimized, and my experiences ignored. When I've named names of people, or even just events, or organizations, it's been even worse.

At best, I've had my work and credentials ignored and belittled, I've had the CEO of a major tech company leave me threatening voicemails. I've had media drag my employer into things I was doing on personal time with unpleasant consequences. I've had death threats, and rape threats, hackers, and stalkers. I've been told I was asking to be assaulted by wearing a skirt to a tech conference, and that I was so ugly I should be grateful for it when it happened.

I've had all of these things not once, but over, and over again. I tell you this again, not to shock you, but to point out the differences in the responses. When people are sending me emails, they're not calling me and ass. The ways that we have all learned to interact as people, as women, as men, as technologists, affect the way we see these issues, and the way that we respond to them. That's why I think it's so important for us all to work together to change things.

People ask me, "What would you recommend to a woman who's wanting to get into open source? How do we build these open source communities that are friendly, and welcoming?" I think the answer is that we have to build them from the ground up. It's wonderful to have people like Schwern going into the existing communities, working with PERL. I do work in Apache. There are people in all of the major open‑source communities.

But my recommendation for anybody trying to build a community nowadays is start fresh. The ADA initiative are working not only on open computing, but on all kinds of open stuff, open culture, open dialogue, open democracy. The outreach program for women, originally run by the GNOME foundation provides paid internships, and again those of you with a sociology background will understand how important that is for minorities and for diversity to be able to get paid for this work.

They have now expanded from being the GNOME foundation, to having internships across a huge range of open source communities and projects, and there are more and more of these communities. There are communities like Dream With, Arc of Our Own, that are already doing great work, and providing welcoming communities, and I hope that we'll see over the next month, six months, year, many more of you joining these communities. I think that's all I had to say, and I think we'll both be happy to take questions.


Michael:  Thank you very much for listening, and for having us here, and I hear you like questions. No?

Audience Member 1:  Hi, thank you both for the excellent talks and insights. I guess I just have some questions that kind of are asking for a little bit more granularity. In your experience, do you find any differences between Europe, which of course is also very diverse, but I've just heard an Irish accent, so I took the opportunity. Europe versus North America on the one hand, and then I was really intrigued by your insight about how a founding dictator, benevolent dictator, might really erect certain barriers, because the project I'm most familiar with is Debian, which doesn't have that.

When they initially proposed Debian woman back in 2004 or '05, they even put it on the bug tracking software, and it was embraced, not really critiqued. I was just wondering if you had all a bit more to say about projects which rely so much on charismatic authority, and whether that always kind of spells some trouble when it comes to diversity?

Noirin:  I'm going to jump in on that. I think you have a very different recollection of the founding of Debian women than I do. Certainly, once they got over the initial shock of, "Why we need this?" Debian did very well, and I think that ties into your first question. I think of Debian as being a more European project. The people who are the core of that community, and although there isn't a single dictator, there is very much a core community.

But the people who were at the core of the community are European, and I think the attitude in Europe, in my experience, and I've only been in the states for two years, but I've been speaking, and on the conference circuit, and working with the Apache Foundation, which is primarily American, for almost 10 years at this point.

I think the tolerance of poor behavior is less in Europe. The tolerance of rude behavior is much greater in the US, and that leads to communities that have this hostile feeling, this ongoing perpetuated problem where we see that the jerks are this huge swath, instead of being a small moderated slice.

Michael:  I'd say I don't have nearly the breadth of experience with projects. I'm just, again, I'm mostly PERL, or world traveling as Noirin does. But I would say that definitely my experience with Europeans is that the concept of American exceptionalism, the idea that the rules don't apply to us, because we're different very, very much comes out in the PERL community, in the American and British dominated PERL community, but...

Noirin:  The one other thing I have to throw in at that point is that Europeans don't have this assumption that, [shouting] "My free speech is what's the most important thing here," which is a very common argument brought up in the open‑source world to anybody who's saying, "Hey, that was a bit out of line."

Michael:  Yeah, "You're censoring me, you're..." Yeah. The other observation is that Americans aren't as much about collaboration as in Europe, because we just don't have to. We don't have to think about working with people with other languages, and other radically different cultures, and maybe people that we were at war with hundred years ago. We can just ignore it.

Noirin:  Broad sweeping generalizations are always a bad idea.

Michael:  Yeah.


Michael:  Yeah.

Audience Member 2:  One of the things that I'm concerned about is that the problems in open source are begetting problems in hiring. Because there's a trend now of hiring managers, taking a resume and going out and looking online to see how active they are in the open source community, which totally discriminates against anybody who, for whatever reason, chooses not to display their work publicly. Essentially, even women who work in proprietary software are going to be at a disadvantage trying to change jobs, because they don't have work out there on GitHub, et cetera.

Do you have any thoughts about how we can kind of counter that practice in these smaller companies that are using open source as the measure of whether or not somebody is worthy as a job candidate?

Noirin:  I will say, I see far fewer laptops open in this room than I typically see at an open source conference, but any of you that have searched for my name will find that I'm actually the first result in Google for my first name alone. It's not necessarily pleasant stuff that's out there. I'm all about people moving away from simply Googling and taking what they see there.

I think it's important to engage in dialogues with candidates. There's a balance to be struck. What can be achieved in an interview or in a day of interviews will only give you such a perspective on the candidate. Having previously worked at Google, where there is this huge focus on putting people through a very rigorous set of interviews and taking only what's happened in those six, forty‑five minute sessions to now working at a very open source company where we do look at people's backgrounds, who they've worked with previously, much more on references and code samples or writing samples, whether those are online or off.

Whether they're publicly available, or whether you have to ask the candidate for them. It's a hard balance to make. I think the key is giving people the opportunity to provide things to a potential employer without having to put them out there for the world. Asking people for samples, I'm a technical writer by profession and its perfectly normal to ask people for writing samples.

For developers, I think open source is still a great way to build up a portfolio, if you don't have one from another job, but it's a tricky question.

Michael:  I don't have anything to say about that.


Audience Member 3:  Hi, right in front of you. There you are.

Noirin:  Thank you.

Audience Member 3:  I'm here actually representing the information security world.

Michael:  All of it?


Audience Member 3:  Essentially now, until we can get more groups to join us, we are the only affinity group that has a direct connection to the information security world and Defcon specifically, and all that. This issue is red hot in that area if you guys are familiar with that. What I'm here for is to figure out that, that community does not respond very well to the direct pointing out of various different situations, as some initiatives have been using, and looking for the tools to speak more directly to them.

I guess my question to you guys is, how do you find that balance between a more direct, we'll say initiative based approach, and a more indirect kind of...I'm going to talk to you on your level, like you provided in the keynote, because that's actually critical to our success in bringing, what we get the tools back to those people in that community.

Noirin:  I don't know how many of you know about the red‑card/yellow‑card initiative at Defcon. One of the long time Defcon attendees essentially got fed up with poor behavior at the conference and printed out the red and yellow cards. The yellow cards saying, "Hey that wasn't really cool, back off a bit", the red cards were "Uh‑uh, way out of line, this is not OK."

These have been going on for two years now, I think? Just one? Though certainly they've been at more than one conference in the security space, but the feedback I have heard has been, again, the insiders, outsiders. The private feedback among the women is almost exclusively positive. People are happy to have it. They're happier to go to events.

There are a small number of very, very vocal women in on the inside who are saying, "No, we don't want any part of this. We don't want to be represented by this." There are large numbers of men on the inside who are saying, "Why couldn't you just have spoken to me?"

I can say as a woman who has stood up and spoken to guys who are crossing the line, "The reason I can't stand up and speak to you is because the last half‑dozen times I was laughed off. I was told I was being unreasonable. I was groped by somebody else. I can't stand up and say it to you directly, because I have and we have, as a collective, this experience of it going poorly." I think you need to have both.

I think you need to have the women empowering themselves, and choosing initiatives that work for them, and you're going to get the backlash from the men saying, that's not cool. You have this privilege going on. Nobody likes to be called on their, pardon me I'm Irish, on their crap.

Nobody likes to be told that they're being a jerk. That's your job. You guys got to level with each other and figure out how you talk to the guys, but I think it's important to allow and empower the women to come up with their own solutions too.

Michael:  Something that I have learned in the process and reason that I point it out in both the bad experience this year, and the bad experience last year, is that a lot of the response that I got to my...I did it in two posts. The response I got to the first post involving the experience this year was, "Why didn't you do more? Why didn't you do this? Why didn't you do that? Why didn't you talk to this person? Why didn't you do more?"

The second post with a point was, "This has all happened before. You're acting like the first reported incident. The first publicly reported incident is the first incident. You don't understand that there have been a whole chain of garbage before this, and this is just the last straw, and going public is the last straw."

I think that to get through to a lot people is that, what you see on is the tip of the iceberg. Why can't you just tell men to stop it? Why do you need these cards? It's to get across that, because there had been so much garbage going on.

The other thing that I would say is that it takes time. This has to be dealt with as an emotional issue, not really a logical issue. You won't really be able to say the magic words to suddenly make people go...although I have encountered people who will like, do that. You can have a very analytical conversation, they're fascinating.

But for the most part the community has to process this. Every year it gets a little better. It's kind of awful during the process. Then the next year it gets moved forward a little bit, and moved forward a little bit. Yeah, we can talk more later.

Audience Member 4:  Hi.

Michael:  Hi.

Audience Member 4:  I really appreciate that you invited people to create something new in the context of abuse, or less than not enjoying. Also, just thank you. It was a moving statement that you did here, and we appreciate that. What do you think keeps people into a particular small Open Source Community?

When, again, we are creative people, so we can be creating new ones. What are some of the arguments or how can you support people to make a move of creating a new community?

Noirin:  I think there's two separate questions there, "Why do we stay?" and, "How can we encourage people to build better things?" Why do we stay? Open Source is incredible. It's absolutely amazing. The computer that beat Jeopardy was powered almost exclusively by Apache software.

The Internet? 50 percent of it is run on our web server. This is something huge, it's brilliant. I don't think any of you in this room aren't running his software somewhere. That is just this amazing thing that we're creating with this technology, and we're creating it as global groups, it's just stunning.

Anything you start fresh is going to start smaller than that. It's never going to have that really amazing powerful pull to a large group of people, and I think it's important to create something that has that powerful pull to you. But to get this whole global community of people from all walks of life, at Apache we have pilots, we have homemakers, we have developers. We have all kinds of people.

I'm sure in PERL, the same, and getting all those people together is just deeply inspiring and wonderful. How we encourage people to build new things? I think is by building them. Is by going out and putting them together. By getting two or three people and saying, "OK we're going to do a Dynamic Languages BarCamp in Portland next Autumn."

Michael:  [laughs] Did you just announce that?

Noirin:  I just announced that.

Michael:  [laughs]

Noirin:  Yeah. We're going to do a Dynamic Languages BarCamp and we're hoping to build up to Dynamic Languages Conference in Portland. I work on conferences, and Schwern is good with dynamic languages, and so that's what we're building. What are you building? What are each of you building?

Please, put them on Twitter. Tell us what you're building. Bring us all in together and grab the people who are interested in building the things that you're interested in, even if it's only two or three people.

Michael:  From my end, what keeps me in is that really the center controls, in this case, in PERL, the language, the infrastructure, the libraries that I have to use. A given PERL program can require hundreds of libraries written by dozens and dozens of different people. There is an interaction. If you want to interact with the community, if you want to patch the language, you have to talk to a mailing list, you can't just go onto GitHub, and submit your patch, and have your pull request, and then be gone.

Unfortunately, at a certain level of involvement with an open source language or project, a major one, you have to be directly involved with the community. I am so personally invested in PERL. It's my friends, my social groups, my job, my hobby, it's still really fun. I can't just abandon it, it's part of my identity. I think there are a lot of people, again who are in a position of power, are kind of stuck to a language, they've built up all this investment in.

Encouraging new projects, part of that, is encouraging, is just encouraging new people so that they can...When somebody, that is not a programmer goes, "I have a great idea," they don't have to go out and find a programmer. They can become a programmer. I was just talking to one of the Udacity folks on being like, "I love your intro to computer science video," because it started from zero, and the project was a serious one.

The project was, build a web search engine. I don't think I could build a search engine, and he teaches you all the way through, in Python, how to do it. That enables people who aren't programmers to do what they want, to make their own projects, do their own work.

Valerie Taylor:  I'm Valerie Taylor, and I'm the Senior Associate Dean of academic affairs in The College of Engineering at Texas A&M University so I loved the "howdy."

Michael:  I'm from New York.


Valerie:  I'm also Executive Director of the Center for Minorities and People with Disabilities in IT, so I want to, on behalf of NCWIT, I want to thank you for having a voice in the insiders, in the open source community, and for being that voice of inclusion, so thank you.



Transcription by CastingWords